A Modified Hybrid Brain-Computer Interface Speller Based on Steady-State Visual Evoked Potentials and Electromyogram.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Human-Computer Interaction

Background:

• Enhancing information transfer rate (ITR) in steady-state visual evoked potential (SSVEP)-based spellers requires more flickering symbols, potentially reducing accuracy.
• Hybrid brain-computer interfaces (BCIs) integrate multiple control signals to improve performance.
• Simultaneous hybrid BCIs utilize concurrent control signals for enhanced ITR.

Purpose of the Study:

• To develop and evaluate a hybrid BCI speller combining electromyogram (EMG) and SSVEP to increase the information transfer rate (ITR).
• To achieve a 36-character selection using only nine stimulus symbols by integrating SSVEP and EMG modalities.
• To assess the performance and usability of the proposed hybrid BCI speller for practical applications.

Main Methods:

• A simultaneous hybrid BCI approach was employed, combining SSVEP and EMG signals.
• Nine stimulus symbols were used, each representing four characters based on EMG muscle activity states.
• SSVEP identified the selected symbol, while EMG determined the specific character within the symbol's options. Frequency encoding was used for EMG, and latency for SSVEP.

Main Results:

• The hybrid system achieved an average ITR of 96.1 bit/min with 91.2% accuracy.
• The speller demonstrated a speed of 20.9 characters/min.
• Electromyogram (EMG) classification accuracy was approximately 100%, while steady-state visual evoked potential (SSVEP) accuracy varied from 80% to 100% across subjects.

Conclusions:

• The proposed hybrid BCI speller offers improved speed compared to existing SSVEP or SSVEP-EMG systems.
• The system provides a user-friendly and practical solution for speller applications.
• The integration of SSVEP and EMG effectively enhances BCI speller performance.